07 Jan, 2011
12 commits
-
Signed-off-by: Ferenc Wagner
Signed-off-by: David S. Miller -
The new firmware interface requires each Slow Path Queue (SPQ) message's
type field to include the function number. The existing code does not
do this consistently. We fix this by OR'ing in the function number
into the type field centrally in cnic_submit_kwqe_16().Signed-off-by: Michael Chan
Signed-off-by: David S. Miller -
In 1ae4de0cdf855305765592647025bde55e85e451, the secctx was exported
via the /proc/net/netfilter/nf_conntrack and ctnetlink interfaces
instead of the secmark.That patch introduced the use of security_secid_to_secctx() which may
return a non-zero value on error.In one of my setups, I have NF_CONNTRACK_SECMARK enabled but no
security modules. Thus, security_secid_to_secctx() returns a negative
value that results in the breakage of the /proc and `conntrack -L'
outputs. To fix this, we skip the inclusion of secctx if the
aforementioned function fails.This patch also fixes the dynamic netlink message size calculation
if security_secid_to_secctx() returns an error, since its logic is
also wrong.This problem exists in Linux kernel >= 2.6.37.
Signed-off-by: Pablo Neira Ayuso
Signed-off-by: David S. Miller -
Since nf_ct_expect_dst_hash() may be called without nf_conntrack_lock
locked, nf_ct_expect_hash_rnd should be initialized in the atomic way.In this patch, we use nf_conntrack_hash_rnd instead of
nf_ct_expect_hash_rnd.Signed-off-by: Changli Gao
Acked-by: Eric Dumazet
Signed-off-by: David S. Miller -
RFC3168 (The Addition of Explicit Congestion Notification to IP)
states :5.3. Fragmentation
ECN-capable packets MAY have the DF (Don't Fragment) bit set.
Reassembly of a fragmented packet MUST NOT lose indications of
congestion. In other words, if any fragment of an IP packet to be
reassembled has the CE codepoint set, then one of two actions MUST be
taken:* Set the CE codepoint on the reassembled packet. However, this
MUST NOT occur if any of the other fragments contributing to
this reassembly carries the Not-ECT codepoint.* The packet is dropped, instead of being reassembled, for any
other reason.This patch implements this requirement for IPv4, choosing the first
action :If one fragment had NO-ECT codepoint
reassembled frame has NO-ECT
ElIf one fragment had CE codepoint
reassembled frame has CESigned-off-by: Eric Dumazet
Signed-off-by: David S. Miller -
Return -ENOMEM instead of 0 for the case of mdiobus_alloc and kmalloc failure.
Signed-off-by: Axel Lin
Acked-by: Florian Fainelli
Signed-off-by: David S. Miller -
The original code has a use after free bug because it's not using the
_safe() version of the list_for_each_entry() macro.Signed-off-by: Dan Carpenter
Signed-off-by: David S. Miller -
There is a "goto nla_put_failure" hidden inside the NLA_PUT() macro, but
we're holding the dcb_lock so we need to unlock first.Signed-off-by: Dan Carpenter
Signed-off-by: David S. Miller -
Return PTR_ERR(port->phydev) instead of 1 if phy_connect failed.
Signed-off-by: Axel Lin
Acked-by: Krzysztof Halasa
Signed-off-by: David S. Miller -
Ethertype used by HPNA control protocols (LARQ, rate, link, etc) and by
Broadcom wlan drivers for local signalling.Signed-off-by: Henry Ptasinski
Signed-off-by: David S. Miller -
Leonardo Chiquitto found poll() could block forever on tcp sockets and
Urgent data was received, if the event flag only contains POLLPRI.He did a bisection and found commit 4938d7e0233 (poll: avoid extra
wakeups in select/poll) was the source of the problem.Problem is TCP sockets use standard sock_def_readable() function for
their sk_data_ready() handler, and sock_def_readable() doesnt signal
POLLPRI.Only TCP is affected by the problem. Adding POLLPRI to the list of flags
might trigger unnecessary schedules, but URGENT handling is such a
seldom used feature this seems a good compromise.Thanks a lot to Leonardo for providing the bisection result and a test
program as well.Reference : http://www.spinics.net/lists/netdev/msg151793.html
Reported-and-bisected-by: Leonardo Chiquitto
Signed-off-by: Eric Dumazet
Tested-by: Eric Dumazet
Signed-off-by: David S. Miller
06 Jan, 2011
10 commits
-
unix_release() can asynchornously set socket->sk to NULL, and
it does so without holding the unix_state_lock() on "other"
during stream connects.However, the reverse mapping, sk->sk_socket, is only transitioned
to NULL under the unix_state_lock().Therefore make the security hooks follow the reverse mapping instead
of the forward mapping.Reported-by: Jeremy Fitzhardinge
Reported-by: Linus Torvalds
Signed-off-by: David S. Miller -
commit 57dbb2d83d100ea (sched: add head drop fifo queue)
introduced pfifo_head_drop, and broke the invariant that
sch->bstats.bytes and sch->bstats.packets are COUNTER (increasing
counters only)This can break estimators because est_timer() handles unsigned deltas
only. A decreasing counter can then give a huge unsigned delta.My mid term suggestion would be to change things so that
sch->bstats.bytes and sch->bstats.packets are incremented in dequeue()
only, not at enqueue() time. We also could add drop_bytes/drop_packets
and provide estimations of drop rates.It would be more sensible anyway for very low speeds, and big bursts.
Right now, if we drop packets, they still are accounted in byte/packets
abolute counters and rate estimators.Before this mid term change, this patch makes pfifo_head_drop behavior
similar to other qdiscs in case of drops :
Dont decrement sch->bstats.bytes and sch->bstats.packetsSigned-off-by: Eric Dumazet
Acked-by: Hagen Paul Pfeifer
Signed-off-by: David S. Miller -
Somehow this snuck into my earlier patch, and
only now did I see a compiler warning:net/mac80211/led.c:218:13: warning: function '__ieee80211_create_tpt_led_trigger' with external linkage has definition
Remove the stray extern.
Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When the driver has remain-on-channel offload,
implement off-channel transmission using that
primitive.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
This allows drivers to support remain-on-channel
offload if they implement smarter timing or need
to use a device implementation like iwlwifi.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
…wireless-next-2.6 into for-davem
Conflicts:
net/bluetooth/Makefile -
Signed-off-by: John W. Linville
-
Signed-off-by: John W. Linville
-
Signed-off-by: John W. Linville
05 Jan, 2011
18 commits
-
Fix test in lbs_spi_thread(). down_interruptible() can return -EINTR, but
not EINTR.Signed-off-by: Roel Kluin
Cc: Dan Williams
Signed-off-by: Andrew Morton
Signed-off-by: John W. Linville -
The key documentation is slightly out of date, fix
that. Also, the list entry in the key struct is no
longer used that way, so list_del_init() isn't
necessary any more there.Finally, ieee80211_key_link() is no longer invoked
under RCU read lock, but rather with an appropriate
station lock held.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
To inject a packet in monitor mode, a dummy station has
to be associated with the monitor interface in the target.
Failing to do this would result in a firmware crash on the device.Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville -
ath5k_eeprom_mode_from_channel() returns -1 on error but we're storing
the result in "ee_mode" which is an unsigned char. This breaks the
error handling. This patch makes "ee_mode" an int.Signed-off-by: Dan Carpenter
Signed-off-by: John W. Linville -
Remove some obvious looking dead code and rename few functions
Signed-off-by: Mohammed Shafi Shajakhan
Signed-off-by: John W. Linville -
This reverts enables the reorder release timer once again.
The issues laid out in:
Have been addressed by:
mac80211: serialize rx path workers
mac80211: ignore PSM bit of reordered framesSigned-off-by: Christian Lamparter
Signed-off-by: John W. Linville -
This patch addresses the issue of serialization between
the main rx path and various reorder release timers.It converts the previously local "frames" queue into
a global rx queue [rx_skb_queue]. This way, everyone
(be it the main rx-path or some reorder release timeout)
can add frames to it.Only one active rx handler worker [ieee80211_rx_handlers]
is needed. All other threads which have lost the race of
"runnning_rx_handler" can now simply "return", knowing that
the thread who had the "edge" will also take care of their
workload.Signed-off-by: Christian Lamparter
Signed-off-by: John W. Linville -
Restart the beacon timers only if the beacon
was already configured. Otherwise beacons timers
are restarted unnecessarily in unassociated state too.Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan
Signed-off-by: John W. Linville -
This reverts commit 0ce3bcfc84900a64347b0fe1140229bd81314008.
Event though with the above commit we obtain the configured DTIM period
from the AP rather than always hardcoding it to '1', this seems to cause
problems under the following scenarios:
* Preventing association with broken AP's
* Adds latency in roaming
So its better to always use the safe value of '1' for dtim periodCc: Jouni Malinen
Signed-off-by: Mohammed Shafi Shajakhan
Signed-off-by: John W. Linville -
Fixes the misplaced article in the following:
"cfg80211: Updating information on frequency 5785 MHz for
20 a MHz width channel with regulatory rule:"Signed-off-by: Bob Copeland
Signed-off-by: John W. Linville -
Fixed a bug where if a mesh interface has a different MAC address from its bridge
interface, then it would not be able to send data traffic to any other mesh node.
This also adds support for communication between mesh nodes and external bridged
nodes by using a 6 address format if the source is a node within the mesh and the
destination is an external node proxied by a mesh portal.Signed-off-by: Joel A Fernandes
Signed-off-by: John W. Linville -
An Rx DMA descriptor can have multiple error bits set, and some error
bits (e.g. MIC failure) are filtered by the driver based on other criteria.
Remove the 'else' in various error bit checks so that all error information
is properly passed to the driver.Signed-off-by: Felix Fietkau
Signed-off-by: John W. Linville -
And add the copyright/license header.
Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville -
When returning to the operating channel, a full HW
reset has to be done instead of a fast channel change.
Since sw_scan_complete() is called after the config() call for the
home channel, we end up doing a FCC. Fix this issue by checking
the OFFCHANNEL flag to determine FCC.Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville -
The device has to be reset when a FATAL event is received.
Not doing so would leave the card in a non-working state.Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville -
There is no need to lock the various work cancellation
calls. This will be helpful when handling FATAL events.Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville -
When doing a channel change, the pending URBs have to be killed
properly on calling htc_stop().This fixes the probe response timeout seen when sending UDP traffic at
a high rate and running background scan at the same time.Cc: stable
Signed-off-by: Sujith Manoharan
Signed-off-by: John W. Linville
